quality assurance in pesticide sampling and analysis · made in sampling and analytical procedures...

Quality Assurance in PesticideSampling and Analysis

Ivan R. Kennedy, Nazir Ahmad, Helen Beasley, John Chapman,John Hobbs, Bruce Simpson, Nicholas Woods

Occasional Paper No 14/98

COTTONRESEARCH ANDDEVELOPMENTCORPORATION

Authors:

Ivan R. KennedyProfessor in Agricultural and Environmental ChemistryCRC for Sustainable Cotton ProductionDepartment of Agricultural Chemistry & Soil ScienceThe University of Sydney, NSW 2006Phone: (02) 9351 3546Fax: (02) 9351 5108Email: [email protected]

Nazir AhmadAustralian Water Technologies Environmental Laboratories51 Hermitage Road, West Ryde NSW 2114Phone: (02) 9334 0725Fax: (02) 9334 0741

Helen BeasleyGrain Quality Research LaboratoryCSIRO Division of Plant IndustryPO Box 7, North Ryde, NSW 1670Phone: (02) 9490 8413Fax: (02) 9490 8419Email: [email protected]

John ChapmanCentre for Ecotoxicology, EPALocked Bag 1502, Bankstown, NSW 2200Phone: (02) 02 9514 4151Fax: (02) 9514 4163Email: [email protected]

John HobbsNSW Department of Land andWater Conservation LaboratoriesGuess Avenue, Arncliffe, NSW 2205Phone: (02) 9597 4444

Bruce SimpsonAgricultural ChemistryQueensland Department ofNatural Resources80 Meiers Road, Indooroopilly, Qld 4068Phone: (07) 3896 9486Fax: (07) 3896 9623Email: [email protected]

Nicholas WoodsThe Centre for Pesticide Applicationand SafetyThe University of Queensland,Gatton CollegeLawes, Qld 4343Phone: (07) 5460 1293Fax: (07) 5460 1283Email: [email protected]

Published by:

Land and Water Resources Research and Development CorporationGPO Box 2182Canberra ACT 2601Phone: (02) 6257 3379Fax: (02) 6257 3420Email: [email protected] Page: www.lwrrdc.gov.au

© LWRRDC

Disclaimer:

The information contained in this publication has been published by LWRRDC to assist public knowledge anddiscussion and to help improve the sustainable management of land, water and vegetation. Where technicalinformation has been prepared by or contributed by authors external to the Corporation, readers shouldcontact the author(s), and conduct their own enquiries, before making use of that information.

Publication data:

‘Quality Assurance in Pesticide Sampling and Analysis’ by Ivan Kennedy, Nazir Ahmad, Helen Beasley,John Chapman, John Hobbs, Bruce Simpson and Nicholas Woods, LWRRDC Occasional Paper No 14/98.

ISSN 1320-0992

ISBN 0 642 26727 8

Typesetting: Mastercraft, Canberra

Printing: Communication Station, Canberra

July 1998

Contents

Specifications for Quality Assurance program1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.3 General principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.4 QA Analytical Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.5 Criteria for acceptability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.6 Reporting of QA information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Recommended protocols—sample collection and storage2.1 Soil sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.2 Water/sediment samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.3 Foliage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.4 Drift samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.5 Bottom sediments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Extraction of field samples3.1 Dichloromethane extraction of water samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.2 Extraction of soil for GLC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.3 Extraction of foliage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.4 Cleanup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3.5 Extraction of soil for immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Gas chromatographic analysis4.1 Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

4.2 Gas chromatographs and chromatographic columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

4.3 Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

4.4 Calibration procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Quality Assurance in chemical sampling and analysis5.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5.2 Progress and conclusions to date . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5.3 Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5.4 Stability and homogeneity of samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5.5 Analysis of data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

iv Quality Assurance in Pesticide Sampling and Analysis

AppendicesAppendix 1 Sampling designs at the Narrabri and Warren field sites . . . . . . . . . . . . . . . . 22

Appendix 2 Dichloromethane extraction of water samples . . . . . . . . . . . . . . . . . . . . . . . . 23

Appendix 3 Extraction of soil and foliage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Appendix 4 Gas chromatographic analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Appendix 5 Immunoassay of endosulfan and diuron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Figures and TablesFigure 5.1—Soil samples from Emerald (September 1994) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Figure 5.2—Soil samples from Warren (17 March 1995) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Figure 5.3a—Water runoff from Warren (24 November 1994) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Figure 5.3b—Youden plots for analysis of water runoff from Warren (24 November 1994) . . . . . 17

Figure 5.4a—Water runoff from Warren (24 November 1994)Results of the water and sediment fractions analysed separately (C,D). . . . . . . . . . . . . . . . . . 17

Figure 5.4b—Youden plots for above analyses of water runoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Figure 5.5—Water samples from Emerald (December 1994) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Figure 5.6—Drift samples from Warren (21 December 1994) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Figure 5.7—Immunoassay validation of soil samples from Warren (1994–95) . . . . . . . . . . . . . . . . 20

Table 5.1—Initial sample interchange schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Table 5.2—Foliage samples from Warren (1995) analysed for chlorfluazuron . . . . . . . . . . . . . . . . 21

Table 5.3—Endosulfan sulphate (mg kg–1) found in sediment samples . . . . . . . . . . . . . . . . . . . . . 21

Table A.2.1—List of the classes of organic compounds for whichmethod is applicable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Table A.2.2—List of preparative apparatus used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Table A.4.1—List of pesticides for which WELOR301 is applicable . . . . . . . . . . . . . . . . . . . . . . . . 34

Table A.4.2—Nitrogen and phosphorus containing pesticides forwhich WELOR204 is applicable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Quality Assurance in Pesticide Sampling and Analysis 1

Specifications forQuality Assurance program

1.1 IntroductionTo achieve consistency in data sets and to assist in interpretation and comparison of results, at thebeginning of this program, the Land and Water Resources R&D Corporation (LWRRDC) requestedthat common protocols be established for sampling, data production and reduction, validation andcompleteness, for use by the separate projects of the joint program. It was recognised that themethods of analysis vary slightly to suit local needs of each laboratory and project, but a commonbasis was necessary for assessment of data, to provide consistency and for valid comparisons to bepossible. LWRRDC also requested that a Quality Assurance project (QA) be conducted as part of theMinimising the Riverine Environmental Impact of Pesticides R&D Program. Both the originalrecommendations for sampling and analysis and the results of the QA are described in thispublication.

Most of the methods used by the research groups and laboratories involved in the program arebased on recommended procedures of the US EPA or the Australian Standards Association; beingvalidated. These should therefore produce high quality data.

1.2 ObjectivesThe quality assurance objectives involve measurements of data in terms of precision, accuracy,representativeness, completeness and ease of comparison.

The three main analytical laboratories involved in the program (Agricultural Chemistry, QDPI,Indooroopilly, Queensland; BCRI, NSW Agriculture, Rydalmere; NSW Department of WaterResources—now Department of Land and Water Conservation, Arncliffe) had NATA (NationalAssociation of Testing Authorities) accreditation and each was experienced in analysis of the full rangeof the pesticides whose behaviour is to be investigated in the program. Therefore, high standards ofanalysis were expected to be maintained. In cases where analyses were conducted elsewhere, carefulreference to and use of procedures used in these laboratories was expected to be maintained.

In addition, to ensure consistency of data sets throughout the program, a small proportion (ca. 5%)of soil, water or other samples analysed in any laboratory were required to be confirmed byanalyses of the same samples in at least one of the other laboratories participating in the program.Full details of the quality assurance program utilised in this program follow.

The Quality Assurance (QA) program proposed was designed to:

• Control and document the laboratory processes in general use in each of the participatinglaboratories. These are documented as appendices in this Manual.

• Through a program of exchanged samples from the various experimental sites, to conduct ananalytical program to verify the reproducibility and general good quality of the data collectedin each participating laboratory.

• The QA program was administered and assessed by the University of Sydney, a participant inthe research of the joint program but one with an independent role, with relatively minorinvolvement in routine pesticide analysis. The results of the QA program were to be reportedin a manner enabling the end-users of the results of the joint program to assess the data quality.

2 Quality Assurance in Pesticide Sampling and Analysis

1.3 General principlesIt was expected that normal operating procedures in each of the laboratories would providemechanisms for identifying and correcting laboratory errors, as detailed elsewhere in this manual.

The laboratory QA programs were expected to control and document each step of the analyticalprocess. For example, in each of the participating laboratories, it was expected that this would beaccomplished by controlling and documenting:

• Laboratory capability—using established criteria for facilities, staff, equipment, andoperational procedures to ensure that the laboratory has the ability to do quality work.

• Laboratory performance—by monitoring the day-to-day performance of the laboratory throughthe use of control samples to ensure consistency of quality performance.

• Matrix effects—by using matrix spikes, matrix duplicates, matrix spike duplicates, matrixsurrogate recoveries, matrix standard additions, etc to assess the affect of the sample matrix onmethod performance and data quality.

As the main analytical laboratories involved in the joint program are NATA accredited (NSWDLWC,BCRI—NSW Agriculture, QDPI—Indooroopilly), this level of control was expected to beautomatically applied. Other laboratory analyses in the program should ensure similar standards. Inaddition, an integral part of the program was the reporting of sufficient QC information to allow thejoint program as a whole to form an assessment. Any reported information was to be supported byextensive documentation kept on file in the laboratory but available to those concerned upon request.

1.4 QA analytical programSchedules for the interchange of analytical samples were circulated. These schedules catered forthe following aspects:

• Analyses of samples were included for all aspects of the joint program, including soil, waterrunoff including sediments, drift and volatilisation, and foliage. In addition, all projects withinthe joint program were expected to participate by supplying samples of an appropriate naturefor interchange (including rain simulator and ecotoxicology studies).

• An orderly shipment of samples from field sites was ensured by circulated spreadsheetschedules. QA samples were requested to be given priority for analysis, allowing corrections inprocedures, indicated as necessary, to be made as soon as possible.

• An intensive period of sample interchange occurred during 1994 and 1995, and some sampleinterchange was to continue throughout the life of the joint program.

1.5 Criteria for acceptabilityBased on comparisons of data from the various field sites obtained in the first year of the jointprogram, there did not appear to be major anomalies, although significant improvements weremade in sampling and analytical procedures as a result of this experience. In particular, theeffectiveness of methods for analysing runoff samples with a heavy sediment load were examined.


Given the fact that it is well known that repeated analyses of environmental samples for pesticideshave standard errors often about one-third to one-half of the mean value, it was considered asundesirable to be too prescriptive about exact numerical criteria for assessment of QA. However, itis suggested that analyses of the same samples by different laboratories involved in the jointprogram should be expected to overlap in the same range of values in this variance range. Wheremultiple samples have been analysed (eg. of field soil cores), the variance of the data should fall inthe same range for each of the different laboratories.

Standard statistical procedures for reproducibility were to be applied to the data, and the resultsexpressed in a convenient graphical form.

1.6 Reporting of QA informationMost monitoring programs do not specify that any quality control (QC) data be reported with theanalytical data. As a result, it is virtually impossible for the end-user to evaluate the quality of thedata received. One program that is notable for the ‘deliveries’ required is the USEPA/CLP (ContractLab Program). Deliveries specified by this program include all QC information plus all raw datagenerated during the analytical process. While some information above and beyond a table ofanalytical results is needed to evaluate data quality, it is debatable that the average user ofenvironmental data quality, has the expertise to use and evaluate a data package which includes allof the raw data.

As a compromise between these ‘all or nothing’ approaches, a policy of preparing the followinginformation for each data set in annual reports or in the Protocols Manual was recommended forthe joint program:

• analytical results, reported with appropriate significant figures;

• reporting limits for each analyte (substance being analysed);

• method reference;

• results of laboratory control samples;

• results of reagent blanks;

• commentary on any anomalies encountered during the analysis; and

• recoveries of surrogates.

This information was expected to be extensively documented in material kept on file in thelaboratory and available to the end-user upon request. This documentation should include:

• all raw data from environmental and QC samples;

• all calibration data associated with the samples;

• standard operating procedures which define each step of the analytical process (available ateach laboratory—many of these are given in the appendices of the manual);

• descriptions of facilities, equipment and staff of the laboratory; and

• accreditations held by the laboratory.

The results of the QA program were to be used to monitor and improve analytical methods. Inaddition, a report of the QA program was to be made by the University of Sydney, to the annualworkshop, and a final report at the end of the joint program.


Recommended protocols—sample collection and storage

As a result of consultation within the joint program, a set of recommendations were assembled.Their validity was assessed as part of the QA program. The recommendations, includingmodifications made during the program, are presented here

Each sample container should be clearly marked using permanent ink or laser-printed labels withthe following information:

• date and time of collection;

• place of collection; and

• sample type and identification.

A chain of custody procedure would ensure the legitimacy of each sample. Logbooks and samplecollection forms should contain information such as:

• site of sampling;

• date and time sampled;

• sample identification code;

• sample matrix (soil, water, composite);

• treatment such as preservation, if any;

• identity of sampler;

• method of transport;

• destination;

• specific analyses required (if applicable);

• date and time of arrival in laboratory; and

• name and signature of person taking custody.

The analytical laboratory should be advised (by fax) when samples are despatched, indicatingnumber and type of samples and estimated arrival time.

2.1 Soil samplingAll analytical data for soil was to be reported on a moisture-free basis, using dry weight obtained ina forced-air oven on exposed soil sub-samples dried overnight at 60–105°C. Particularly withvolatile pesticides such as endosulfan, care should be taken to prevent losses by using soil samplesfor extraction with minimum preparation.

Thorough mixing of soil, preferably at the time of sampling to ensure homogeneity for sub-sampling, was to be carried out on a stainless steel or glass tray using a clean or disposable spatula.Equipment can be cleaned by brushing, washed with detergent and tap water, rinsed in acetone orisopropanol, and rinsed three times with clean tap water and allowed to air dry or by use of a cleanpaper towel. Alternatively, soil samples could be mechanically mixed in the laboratory usingequipment such as the Robot Coupe Processor R301 Ultra with stainless steel bowl.


2.1.1 StorageSoil samples are best stored in airtight solvent-washed or brand-new glass jars, verified as pesticide-free, sealed with aluminium or Teflon foil liner, fitted with new plastic screw-caps (see AustralianStandard 2031.1, 1986). Jars should be transported to the field with caps fitted, to minimise thepossibility of contamination. Field samples should be cooled to 4°C or less as soon as possible fortransportation and then held in a deep-freeze until it is possible to perform solvent extraction. Theuse of polythene bags for storage at ambient temperature, of material freshly sprayed with pesticidessuch as endosulfan, is not encouraged, unless samples can be extracted immediately.

2.1.2 SamplesEach soil sample should be selected from a bulked composite of at least 10 well mixed cores, toreduce the sample variance. Depth will depend on the purpose of sampling and the nature of eachchemical. For estimating total soil burden of a chemical not subject to leaching (such asendosulfan), either 5 or 10 cm deep cores would be appropriate, or to cultivation depth. For greatersensitivity in detecting recently applied chemicals, 2.5 cm samples taken carefully with a calibratedtrowel, would be preferable. Sampling depths must be specified. To assist in comparisons betweensites, presentations of residue data as a total burden per unit area of cotton fields (g/ha) should beprovided as an alternative to giving concentrations of residues in soil in mg/kg (ppm).

It could be appropriate on cotton fields to recognise at least three classes of cores, from the tops (T)of beds adjacent to cotton plants, from the edges (E) of irrigation furrows (more prone to erosion)and from the base of furrows (F), where eroded soil may accumulate.

2.1.3 Sampling designSelection of soil cores on cotton fields should be based on a properly justified sampling design.While the nature of this design will depend on the factors to be examined, wherever possible anaccepted sampling design suitable for statistical analysis should be used. Stratified designsfavoured at the Narrabri and Warren sites, are described in Appendix 1. Based on the degree ofuniformity indicated by statistical analyses of data from such designs, it was later considered that azigzag transect of at least 200 metres, with random sampling of 10 cores for preparing a well-mixedcomposite sample, would be adequate. For each cotton field of ca. 50 ha, a minimum of five suchcomposite samples was found to be required.

Sufficient samples should be taken to establish spatial variability in pesticide residues. Samplestaken for such purposes should be clearly labelled, indicating location in the field, whether fromtop, the particular edge or bottom of furrow, and depth.

2.2 Water/sediment samplesExtraction was to be performed on water or turbid water to obtain total pesticide content, or afterremoval of particulate matter by filtration (GFA, 1.2 mm glass membranes) or centrifugation at2,000 rpm in glass vessels, with separate extraction of water (soluble or colloidal fraction) andsediment where sediment loads are significant. The sediment load should be determined on allwater samples, to establish the eroded fraction. The concentration of pesticide residues should begiven as mg/L (ppb).


2.2.1 StorageWater samples of 1 L should be stored in solvent-washed or brand-new (amber) glass bottlesverified as uncontaminated, sealed with aluminium foil or Teflon, fitted with new plasticscrew-caps (see Australian Standard 2031.1, 1986) and chilled immediately to less than 4°C in arefrigerator. Organic solvent (eg. dichloromethane) can be added immediately where convenient tolimit volatilisation or hydrolysis, although care to prevent leakage is essential. Extraction of watersamples with organic solvent should be made within 48 hours and immediately on receipt. Evenso, it can be anticipated that samples containing endosulfan isomers will lose chemical byvolatilisation if jars are not properly sealed, ideally with Teflon (Guerin and Kennedy, 1992), or byhydrolysis if the pH of the water is above 8. Freezing of samples for longer-term storage may also bedesirable where facilities are adequate, provided bottles are only half-filled to prevent breakage andthere are no other problems. The half-life of endosulfan by chemical hydrolysis to endosulfan diolin river water at pH 8.5 is claimed to be less than two days at ambient temperatures (Peterson andBatley, 1993), and it may be desirable to adjust pH to below 7 by addition of phosphate buffer (pH6) or acetate buffer (pH 5.4) to 1–5mM, even for cold storage.

Although a number of pesticides were examined during the joint program, endosulfan was the mainchemical examined particularly at the early stages of the research. For that reason, the QA programfocussed on endosulfan since it was being analysed by each laboratory involved and it wasconsidered that this chemical provided a convenient model for other pesticides.

2.3 FoliageMethods are required to estimate the retention of pesticides on cotton plants, as well as to provideestimates of that washed off in rain and losses by others means such as volatilisation.

The extraction of plant or vegetable material may present special problems, related to the range ofchemicals including pigments that occur in plant tissues. Tissues are normally homogenised toimprove the efficiency of extraction using acetone-hexane or acetone-dichloromethane and clean-up is achieved on a silica column, such as Florisil. A decision whether results will be expressed asmg/kg or per unit area needs to be made, so that the surface area of a sample weight of leaves usedfor extraction can be estimated using a suitable planimeter.

2.3.1 StoragePlant material may metabolise pesticides such as endosulfan rapidly, so it is important to extractleaves as soon as possible, or to freeze samples immediately before extraction later. Substantialquantities of endosulfan sulphate are formed in cotton leaves from endosulfan formulations withinseveral days (Coleman and Kennedy, 1993).

2.4 Drift samplesDown-wind samples in drift studies (Centre for Pesticide Application Safety, C-PAS, University ofQueensland, Gatton) were collected on mobile 10 m telescopic towers fitted with vertical copperwires, aluminium cylinders, nylon gauze, or pipe cleaners, or on filter papers of ca. 200 cm2

attached by elastic bands to aluminium tables supported on wooden dowelling at crop height.Wires were cut into appropriate lengths, wound on aluminium rods and samples stored in glasscontainers (eg. McCartney bottles) sealed with aluminium foil. All collector media were stored bythe same method and Nanograde acetone was added (5–10 mL) as soon as possible to prevent lossesby volatilisation (of endosulfan), although immediate storage of samples held in the deep freezewas also effective.


2.5 Bottom sedimentsCorers capable of sampling sediments under water were developed at the NSW EPA Centre ForEnvironmental Toxicology (J. Chapman). An effective method of sampling sediments is by using aPVC tube fitted with a screw-cap seal following insertion of the tube in sediments under water(S. Kimber, University of Sydney).


Extraction of field samples

All solvents must be verified as contaminant-free prior to use, equivalent to Nanograde(contaminant-free solvent).

3.1 Dichloromethane extraction of water samplesAt the Biological and Chemical Research Institute (BCRI), NSW Agriculture, water samples wereextracted two or three times with dichloromethane, the extract concentrated and the samplecleaned up on an alumina column. The eluate, diluted in hexane, was gas chromatographed on a30 m capillary column and quantitated by electron capture detection.

The method used at the NSW Department of Land and Water Conservation laboratory at Arncliffe(laboratory code WELOR101) is based on USEPA Method 3510, ‘Separatory Funnel Liquid-LiquidExtraction’, for the isolation and concentration of organic compounds from test water samples.

The sample, either as received or pH adjusted, was carefully extracted three times withdichloromethane (in a separating funnel). The extract, dried with anhydrous sodium sulphate, wasthen concentrated using a Kuderna-Danish flask fitted with a 10 mL graduated tube and a three ballSnyder Column. The solvent was exchanged to the specific solvent required for either the analysisor the clean-up stage.

Full details of the scope of use, reagents, apparatus, precautions and procedures are given inAppendix 2.

3.2 Extraction of soil for GLCExtraction of pesticides from soil requires more polar solvents than hexane or dichloromethanealone. A mixed extracting solvent with added acetone, or the use of methanol as the primaryextracting solvent, provides improved extraction of residues. The effectiveness of whicheversolvent is used should be verified by spiking of soil with pesticide standards and by using surrogateanalytes such as dibutylchlorandate or octachlorobiphenyl, to estimate extraction efficiency andsubsequent handling of extracts in the laboratory.

A procedure, developed using acetone-dichloromethane, in the Department of AgriculturalChemistry and Soil Science, University of Sydney, for analysis of endosulfan residues includingendosulfan sulphate, endosulfan ether, endosulfan hydroxy ether and endosulfan diol in cotton-growing soil (Kimber and Kennedy, in preparation), is fully described in Appendix 3.(Dichloromethane has the advantage of non-flammability, although its characteristics as aorganochlorine should be respected). The procedure has been optimised for recovery of thesecompounds (better than 80% of recovery of each), while allowing the removal of several substancesthat would interfere with gas chromatographic analysis by electron capture detection of peaks,particularly when using packed columns but also with capillary columns.

3.3 Extraction of foliageThe extraction of foliage should be by blending with a mixture of dichloromethane and acetone(80:20) in an omnimixer or by soxhlet extraction. Extraction efficiencies should be established byspiking foliage samples at four different levels, one very close but slightly above the detention limitof the method with replication to establish the variation of the method.


The procedure employed at the BCRI, NSW Agriculture, is as follows. Place leaf sample intoextraction thimble within an appropriate concentration of surrogate standard. Extract via soxhletextraction using 25 mL of dichloromethane:methanol (80:20) for at least four hours. Concentrate to5 mL and exchange the solvent to n-hexane. Transfer concentrated extract to 18 cm plastic(polythene) membranes and shake for 24 hours in 100 mL cyclohexane. Concentrate to 5 mL. Pass2ml aliquot through 1g Al2O3 column and elute with 10 mL of 8% acetone in hexane solution.Submit for GC analysis at a final volume of 25 mL containing an appropriate concentration ofinternal standard.

Another method using acetone-hexane as a solvent (see Appendix 3) was extensively adapted fromone developed by Chopra and Mahfong (1977) for extraction of endosulfan and metabolites fromtobacco leaves. It has been employed for study of the rate of dissipation of endosulfan from cottonleaves (Coleman and Kennedy, 1993).

3.4 CleanupBatches of solid supports (eg. alumina, Florisil, etc.) should be pre-assessed for their ability toretain polar compounds and to allow elution of pesticide residues. The cleanup system used forremoving coextractives from extracts should be calibrated by eluting mixtures of standards ofinterest through the cleanup columns and the elution pattern and volume recorded. Each time anew batch of alumina or silica gel is made the elution pattern and volume should be redetermined.

3.5 Extraction of soil for immunoassays (ELISA)An immunoassay method for the detection of endosulfan in water and soil has been developed (Leeet al., 1995; Lee et al., 1997). Soil (10 g) is weighed moist into a glass jar, 25 mL of 90% methanoladded and the sealed sample shaken and stored overnight. Endosulfan content is then estimated on100 µL of the extract following a minimum dilution of 1/100 in water (Lee et al., 1997). The dryweight of soil is measured following thorough air and oven drying of the sample on removal of thesolvent. More details are found in Appendix 5.


Gas chromatographic analysis

4.1 PrefaceAll aspects of good laboratory practice are expected to be in force in analytical laboratories. Allsample handling, extraction, cleanup, analysis, chromatography and confirmation should bereferenced to the laboratory manual and standard operating procedures (SOPs). In reporting results,the precision and accuracy limits as defined in the standard operating procedure of a methodshould be maintained. Duplicate analyses with more than 15% coefficient of variation should berepeated for that sample. Data on precision and accuracy should be provided for each matrix byreplicate analyses of spiked samples just above the detection limits of the method and at three otherspike levels for establishing linearity of each analyte. All negative results should be reported asbelow these established detection limits rather than as not detected.

4.2 Gas chromatographs and chromatographic columnsAnalytical instruments should be maintained as specified by manufacturers. Laboratory personnelresponsible for performing such preventive maintenance should record this in log books availablefor inspection. No sample analysis can be conducted with instruments with not meetingperformance specifications.

Each analysis should be confirmed on columns of different polarity, reporting both sets of results.Ideally, a split injection system with two columns fitted to separate detectors may be used to ensureverification of the identity of peaks.

It is anticipated that analysis will usually be performed on capillary columns, thus reducingproblems of sample cleanup. However, a packed column providing a separation of all theendosulfan breakdown products was described by Guerin et al. (1992).

4.3 StandardsCalibration standards, like internal standards, should be prepared from certified standard materialand kept no longer than six months with new preparations being evaluated against the oldstandard. All standard solution compositions, concentration and lot numbers should be recorded inthe log books.

Every tenth field sample should be analysed in duplicate and each batch of ten should have amatrix blank, a reagent blank and one spike recovery at 0.1 mg/g of analytes of interest, such asa-endosulfan, b-endosulfan and endosulfan sulphate.

4.3.1 Matrix spike samples and blanksMatrix spike and duplicate matrix samples shall also be analysed and results of these analysissubmitted with each batch of 10 or 20 samples. Blanks should be analysed at the frequency of onein 20 samples. All the blanks should be fortified with surrogate compounds.

4.3.2 Internal standard and calibration standardsAll analyses performed on GC/MS will require internal standards for analyte quantisation. Internalstandards shall be prepared from certified reference standards. No solution should be kept longerthan six months. All new preparations of the internal standard solution need to be verified againstthe old solution to ensure acceptability and records kept in the standard log book.


4.3.3 Surrogate and matrix spike standardsAll surrogate and matrix spike standards should be prepared from certified reference standards andcontrol charts for recoveries should be maintained. Information regarding analyte concentration,composition and lot numbers should be recorded for traceability purposes. No solution should bekept longer than three months. New preparations of surrogate and matrix spike solution should beanalysed prior to use to ensure correct concentration levels and absence of any contaminants. Theconcentration of surrogates should be approximately 0.2ppm in solid matrices and 1 ppb in liquid.

4.3.4 Data analysis, validation and reportingPreliminary data reduction performed in laboratories in this program should include analyteidentification, contaminant or interference identification, elimination of false positives and manualquantisation of analyte concentration (eg. GC/MS) when peak splitting occurs. The separation ofchromatographic peaks should be to the baseline and the minimum height of the reported peakshould be at least 10% of the total chart deflection. Peak heights of standards and samples shouldnot exceed 15%.

Analyses of blanks should demonstrate that reagents and glassware are clean and interference-free.Data from chromatograms must be collated and reviewed by a qualified chemist for each set ofsamples. Raw data relating to these samples should be examined for consistency with calculatedresults and spot checks on some of the calculations performed manually. Summary reports shouldbe signed by the chemist and made available for inspection by the supervisor.

4.4 Calibration procedures4.4.1 InstrumentsAll instruments used in the production of data for these projects should be calibrated according toprocedures as prescribed in the SOPs of each method. For example, balances should be calibratedat least once according to manufacture’s specifications and the calibration log book maintained foreach balance. Class S weights should be used to verify electronic calibration.

4.4.2 GC/ECDGas liquid chromatograph should be calibrated at four levels of standard, one very close but above thenoise level of the instrument and then at intervals of 10 times for each analyte. Retention timewindows for each analyte should be established at base line separation. These retention times shouldbe established before each batch of analysis. Instrument log books shall be maintained and willcontain information regarding the usage and maintenance, problems and corrective actions. Prior toany sample analysis the GC operating condition must be maintained so that p,p’-DDT has a retentiontime of between 23–28 minutes on a DB_5 capillary column 30 metres long and of 0.32 mm internaldiameter. Following this adjustment the retention time windows of each analyte must be determined.Calibrated GC operation mandates the analysis of various standards before sample analysis. Thecalibration will remain valid for 72 hours if the individual component pesticide standard meetacceptance criteria during this period. Recalibration will be mandatory after 72 hours.


4.4.3 GC/Mass SpectrometryPrior to any standard, sample, or blank analysis the GC/MS system hardware must be tuned to meetthe ion abundance criteria for the analysis of two reference standards such asdecafluorotriphenylphosphine and p-bromo-fluorobenzene. Then a calibration derived from 20, 50,80, 120 and 160 ng shall be utilised to assess the concentration of semi-volatile compounds. Oncethe system is calibrated the calibration must be verified (tuned and 50 ng standard analysis) onceper each 12 hour period of the operation. Recalibration will be required if the 50 ng standard failsto meet acceptable criteria.


Quality assurance in chemical samplingand analysis

5.1 Objectives1. To control and document the laboratory processes in general use in each of the participating

laboratories. This will be documented in a manual on sampling and analytical protocols. It isanticipated that the manual will be gradually improved up to the end of the joint program, butan advanced version of the manual is currently available.

2. To conduct an analytical program to verify the reproducibility and general good quality of datacollected in each participating laboratory, by analysis of exchanged samples from the variousexperimental sites.

3. To administer the QA program and report the appropriate information in a manner enablingthe end-users of the results of the joint program to assess the data quality.

5.2 Progress and conclusions to dateEach of the main analytical laboratories involved (BCRI (N. Ahmad) Rydalmere, NSW Agriculture;Agricultural Chemistry (B. Simpson) Indooroopilly, QDPI; Arncliffe NSW Water Resources(A. Awad) has provided information on laboratory capacity and performance, including methods ofquality control (matrix effects, ‘spikes’, etc.). Inspections were made of each laboratory, standardlaboratory protocols and primary data generation examined. This information is includedelsewhere in this manual (eg. see Appendices).

An interchange program was successfully conducted between all three main laboratories. Thisresulted in improved procedures for sampling, storage and analysis, particularly considering thevolatility of endosulfan. This program has verified that variations between these three laboratories isquite small (CV of mean values = 10–15% only), so that strong confidence in the accuracy of the datais justified. This degree of variation is much lower than in that found in most comparisons betweencommercial laboratories. In addition, immunoassays being arranged by the University of Sydney (CRCfor Sustainable Cotton Production) and performed by CSIRO Plant Industry, have been included inthis program and excellent agreement between chemical analyses and immunoassays has beendemonstrated.

5.3 ScheduleThe schedule used for the interchange of analytical samples is given in Table 5.1. It catered for thefollowing aspects:

1. Analyses of samples included from all aspects of the joint program, including soil, waterrunoff including sediments, drift and volatilisation, and foliage. In addition, all projects withinthe joint program participated by supplying samples of an appropriate nature for interchange(including rain simulator and ecotoxicology studies).

2. It was expected that QA samples were given priority for analysis, to allow any corrections inprocedures, indicated as necessary, to be made as soon as possible.

3. An intensive period of sample interchange occurred in October–December 1994 to prove thequality of current procedures.


Some variations to the original schedule were made after the first analyses proved to be veryconsistent in their results. Thus, only a few samplings marked with one asterisk in Table 5.1 wereperformed using standard analytical procedures (GLC/MS), and those marked with a cross werecompared for both GC and immunoassay analyses. Some sample interchange continued throughoutthe life of the overall joint program.

Table 5.1Initial sample interchange schedule

No. Date Sample (No.) Source Analyses Laboratories

1* 1993–94 Soil (6) Emerald 12 BCRI, QDPI

2 Nov-94 Runoff (6) Emerald 12 BCRI, QDPI

3 Soil (6) Emerald 12

4+ Rain simulator (6) Warren 12 BCRI, QDPI

5 Dec-94 Runoff (6) Warren 12 BCRI, QDPI

6* Runoff (6) Warren 12 BCRI, DLWC

7 Soil (6) Warren 12 BCRI, QDPI

8+ Drift (5) Warren 10 BCRI, Gatton

9* Jan-95 Runoff (6) Emerald 12 BCRI, QDPI

10 Soil (6) Emerald 12 BCRI, QDPI

11 Foliage (5) Warren 10 BCRI, QDPI

12 Feb-95 Runoff (6) Warren 12 BCRI, QDPI

13* Soil (6) Warren 12 BCRI, QDPI

5.4 Stability and homogeneity of samplesIn conducting QA studies it was important that the analyte be kept stable under likely transport andstorage conditions. Soil and water samples were stored in airtight brand new glass jars, sealed withaluminium foil or Teflon sheet (Ludowici), fitted with new plastic screw caps (see AustralianStandard 20331.1, 1986), and chilled to less than 4°C as soon as they were taken. The samples wereplaced in an insulated box for transportation, and then held in a deep-freeze until the solventextraction was performed. Dispatch of samples was made by express courier.

To ensure soil samples homogeneity, soil cores were thoroughly mixed and then split into two setsof samples, one for each laboratory. This step was of fundamental importance, without which QAresults between laboratories could not be compared.

5.5 Analysis of dataStandard statistical procedures were used to compare the results form the different laboratories,and the results are expressed in a convenient graphical form. When comparing two types ofanalyses on the same sample, eg. GC and immunoassay, a correlation was calculated. When morethan two laboratories were involved in the analysis an ANOVA was applied to the data. In the casefor each pair of results (Youden pairs), the sum and the difference were tabulated. The sum valuereflects the between-laboratory reproducibility of a participant’s results compared to the otherlaboratory. The difference value reflects how well the within-laboratory repeatability compares.


5.5.1 Soil samplesSets of six soil samples from Emerald and five samples from Warren were analysed at QDPI andBCRI and the results are shown in Figures 5.1 and 5.2. There was agreement for total endosulfanresults and for each of the components (alpha, beta and sulphate), and the coefficient of variationfor the replicated Emerald set was acceptable (22.9%), but slightly lower values were consistentlyobtained with this set by the BCRI. However, this discrepancy can be explained because thosesamples were stored for several weeks, partly at ambient temperature, before being sent to thatlaboratory, and the data are consistent with some volatilisation, particularly of the alpha-isomer. Inthe other set (Figure 5.2), the soil samples were analysed more promptly by both laboratories andno consistent difference between the laboratories was obtained.

Both endosulfan isomers, alpha and beta, showed more variable results than those of sulphate,confirming the difficulty in accurately analysing these volatile components in soil samples. Theoverall variation found among soil samples is higher than that of other type of samples, about 15%in the best case.

Figure 5.1Soil samples from Emerald (September 1994)

R2 = 0.98CV (%) = 22.9

Co

eff

icie

nt o

f va

riat

ion

(%)

100

80

60

40

20

0alpha beta sulphate TOTAL

Endosulfan

0

End

osu

lfan

(m

g/kg

)

1

2

3

4

5

6

Samples

A B C D E F

QDPIBCRI


Figure 5.2Soil samples from Warren (17 March 1995)

Figure 5.3aWater runoff from Warren (24 November 1994)

R2 = 0.83CV (%) = 14.8

Coeff

icie

nt o

f va

riat

ion

(%)

100

80

60

40

20

0alpha beta sulphate TOTAL

Endosulfan

0.00

End

osu

lfan

(m

g/kg

)

0.05

0.10

0.15

0.20

0.25

0.30

Samples

QA1 QA2 QA3 QA4 QA5

QDPIBCRI

AB (water & sediment together)CD (separately)

Co

eff

icie

nt o

f va

riat

ion

(%)

alpha beta sulphate TOTAL

EndosulfanSamples

AB CD

QDPIDLWCBCRI

100

80

60

40

20

0

50

40

30

20

10

0

P<0.05


Figure 5.3bYouden plots for analysis of water runoff from Warren (24 November 1994)(i) Extracting the entire sample, water and sediment, together (A,B);(ii) Extracting water and sediment fractions separately (C,D)

Figure 5.4aWater runoff from Warren (24 November 1994)Results of the water and sediment fractions analysed separately (C,D)

50

40

30

20

10

050403020100

(ii)

BCRIDLWC

QDPI(p<0.05)

50

40

30

20

10

050403020100

(i)

DLWC

BCRI

QDPI

QDPIDLWCBCRI

water fractionsediment fraction

water sediment alpha beta sulphate TOTAL

EndosulfanSamples

50

40

30

20

10

0

Co

eff

icie

nt o

f va

riat

ion

(%)

100

80

60

40

20

0


Figure 5.4bYouden plots for above analyses of water runoff

5.5.2 Water runoff samplesTwo pairs of water runoff with a moderate sediment load from rain simulator studies conducted atWarren were analysed by all three laboratories. The results are shown in Figures 5.3 and 5.4.Thelaboratories were asked to analyse these samples using two procedures: (I) the standard EPAmethod simply of three dichloromethane extractions in a separating funnel, followed bycombinations of the extracts (samples A and B); (ii) with a preliminary separation of the sedimentby filtration or centrifugation, followed by extraction of the water by the EPA method andextraction of the sediment separately (samples C and D).

The first procedure showed a significant difference between the results of QDPI and the other twolaboratories, for both alpha and beta isomers as well as for the total endosulfan. This discrepancy,however, disappeared when the laboratories used the second procedure (Figure 5.3, ii). In thiscase the results for water fraction are similar for all three laboratories, except those of endosulfansulphate which varied considerably. The same problem was found in the analysis of the sedimentfraction (Figure 5.4). It seems that endosulfan sulphate is harder to extract reproducibly fromwater samples than the two isomers, contrary to what happens in soil, no matter what procedureis used by the laboratories. This might explain the relatively higher variation found betweenthe three laboratories (22.4% and 27.6%), although the overall agreement of the mean valueswas satisfactory.

These comparisons (and others) suggest that runoff (or river) samples can be analysed quite accuratelyby the standard EPA method provided the sediment load does not significantly exceed 0.1% (w/v).Thus, the routine analysis of river samples by this EPA method (as used by the NSW Department ofLand and Water Conservation), provided it is conducted carefully, yields accurate values. However,for runoff in storms or simulated storms (sediments of 1% (w/v) or greater) separation beforehand isprobably essential for adequate recovery of residues from the sediment fraction. A second test withthree pairs of water samples from Emerald analysed at the Queensland Department of NaturalResources (Brisbane) and NSW Agriculture in Sydney showed better results, with a variation of only10.4% (Figure 5.5). Again, most of the variation was found in the alpha isomer.

DLWC

BCRI

QDPI

50

40

30

20

10

050403020100

50

40

30

20

10

050403020100

sediment fractionwater fraction

DLWCBCRIQDPI


Figure 5.5Water samples from Emerald (December 1994)

5.5.3 Drift and volatilisation samplesA set of 10 drift samples were compared using two analytical procedures, immunoassay and GC, forresidue analysis. The results showed a good agreement (R2 = 0.77, Figure 5.6), though immunoassayanalysis gave higher values than GC analysis (about 10%). The variation within immunoassayresults is larger than that of GC, but this could be due to the smaller quantity of analyte used by thistechnique. In other comparisons of volatilised endosulfan trapped in water trays, using both GCanalysis and immunoassay, the trends of residue concentrations were established to be very similar(data not shown here).

Figure 5.6Drift samples from Warren (21 December 1994)

R2 = 0.96CV (%) = 10.4

QDPIBCRI

Co

eff

icie

nt o

f va

riat

ion

(%)

0

5

10

15

alpha beta sulphate TOTAL

Endosulfan Samples

T2 SELMA CRUMP

100

80

60

40

20

0

GC analysis (R2 = 0.89)Immunoassay (R2 = 0.5)CV (%) = 33.3R2 = 0.77

End

osu

lfan

(m

g/sa

mpl

e)

1.0

0.8

0.6

0.4

0.2

0.01086420

height above ground (m)

1.2

1.4

12 14 16 18 20 22


Figure 5.7Immunoassay validation of soil samples from Warren (1994–95)

5.5.4 ImmunoassaysImmunoassays are being widely applied to this research program (Kennedy et al., 1997). Qualityassurance for this technique has been established by comparing analytical data obtained by gaschromatography and that obtained by immunoassays. It needs to be understood that the endosulfanimmunoassay provides the sum of the toxic forms (ie. both isomers and endosulfan sulphate), withslightly differing sensitivities for each form. However, excellent agreement is being obtained (seeFigure 5.7 for data obtained with soil samples, R2=0.94) between the two methods, verifying thevalidity of the immunoassay technique. Provided water samples are extracted in the field or aretransported chilled very promptly and extracted immediately on receipt, excellent agreement isalso being obtained with runoff samples (Lee et al., 1997; M. Silburn, N. Ahmad, pers. com.).

5.5.5 Foliage samples—HelixMethods for Helix (chlorfluazuron) analysis are poorly developed. No reliable method using gaschromatography is currently available. HPLC methods, however, produce satisfactory results—although the limits of detection are at least ten times less sensitive. A replicated set of six foliagesamples from the fieldsite at Warren were analysed both by AnalChem Bioassay (Lilyfield, NSW) aswell as at the BCRI at Rydalmere and the data communicated independently to the University ofSydney. As shown in Table 5.2, satisfactory agreement between the two laboratories (apart from onesample) was obtained.

R2 = 0.94N = 22

Imm

uno

assa

y (p

pm)

GC (ppm)

1.2

0.9

0.6

0.3

0.0

1.5

0.0 0.3 0.6 0.9 1.2 1.5


Table 5.2Foliage samples from Warren (1995) analysed for chlorfluazuron.

Chlorfluazuron (mg/kg)

Sample ID AnalChem BCRI

A 15.17 13.65

B 4.06 2.06

C 3.08 2.46

D 10.49 3.09

E 5.73 3.57

F <0.1 <0.005

5.5.6 Other laboratories—sedimentsQuality assurance data was obtained for the University of Queensland (Gatton, N. Woods) usingfield samples obtained at Warren; these were also analysed at the BCRI, and by immunoassays/MassSpectrometry at the University of Sydney. Data was also requested from projects on theecotoxicology of endosulfan and pyrethroids conducted by NSW EPA (J. Chapman/R. Hyne). Thelatter program does not involve large numbers of field samples, but it was desirable that a similarlevel of quality assurance also be exercised here. Validation of the analytical data on pesticideresidues in sediments of the ecotoxicology project was obtained from analysis conducted by theNSW Department of Land and Water Conservation laboratory at Arncliffe and the EPA laboratory atLidcombe. In Table 5.3 an example of comparisons for endosulfan sulphate in sediments for dataobtained from the two laboratories is shown—apart from one sample (possibly a false positive)there was good agreement.

Table 5.3Endosulfan sulphate (mg kg–1) found in sediment samples

Sample ID EPA Lab NSW L&W

2 I 8.0 <2

3 II 3.5 4

5 II 8.5 6

6 II 9.1 10

6 III 8.0 9

Pian Creek 4.4 4

5.6 ConclusionAs a result of the QA program and discussions arising from it, improved procedures have beenadopted in all the participating laboratories. In general however, no significant disagreement inanalytical data has been observed in comparisons between the laboratories.

It is concluded that the data being obtained in all laboratories examined in the joint program is bothaccurate and sufficiently precise. Thus, the conclusions from the research program are based ondata assured to be of high quality. The methods used here may also prove to be useful to otherresearch workers interested in the environmental fate of pesticides.

The results of the QA program continue to be used to monitor and improve methods. In thismanual, presented by the end of the joint program by the University of Sydney, QA data from allthe laboratories involved is presented, as part of an upgraded protocols manual.


APPENDIX 1

Sampling designs at theNarrabri and Warren field sites

1. The NSW field site on the Auscott property at Narrabri, Field 21 (1993–94 season) was markedinto 90 approximately square strata each of 80–90 m sides, with surrounding buffer zones. Forthe 1994–95 season, Field 4 on the Auscott property at Warren was marked into 18 squarestrata of 150 m sides. In both sites a sub-sampling design for the strata has been selected (seediagram) to allow representative values and spatial variation to be assessed statistically.

2. As indicated above, soil cores can be collected from the tops, edges of planting beds or in thefurrows.

3. Initially, to reduce the variance of soil samples, composite samples of 20 soil cores werecollected, using a well-mixed sub-sample of each core representative of its full depth.Subsequent statistical analysis indicated that a composite of 10 soil cores was sufficient.

4. Soil cores were selected at regular intervals (alternate beds) on a V-intersection of each stratumat the Narrabri site, or at random lines across each stratum at the Warren site (see diagrams).

5. It is anticipated that soil analysis will be less sensitive to spatial trends than alternativeprocedures such as analysis of runoff water and sediments. However, it is considered essentialto perform both kinds of analyses to establish correlations and seasonal and managementpractices.

A B C D E F G H I1

2

3

4

5

6

7

Head ditch80m

90m

NSW FIELD SITE1993–94

AUSCOTT 21

Sampling pattern

Selected stratum Tail drain Drop box

Æ

8

9

10

NSW FIELD SITE1994–95

AUSCOTT 4

Headditch

Tail

drai

n

Drop box

A

B

C

D

E

F

3 2 1

150m


APPENDIX 2

Dichloromethane extraction ofwater samples

A.2.1 ScopeThe procedure is suitable for the extraction of the classes of organic chemicals listed in Table A.2.1

Table A.2.1List of the classes of organic compounds for whichthe method is applicable.

Chemical Class Analytical Method Extraction pH* Final Solvent

WELOR101

Organochlorine 201 A, B or C Hexane

Polychlorinated 202 A, B or C Hexane

Biphenyls

Pyrethroid 203 A, B or C Hexane

Organo-Nitrogen and 204 B Hexane Phosphorus

Organophosphates 205 B Hexane

Organosulphur 206 B Hexane

Pesticides by GC/MS 301 B Dichloromethane or Hexane

Phenols 302 C Propan-2-ol

Polycyclic Aromatic 303, 402 A, B or C Dichloromethane, Hydrocarbons Acetonitrile or Hexane

* The following sample pH requirements are dependant on the mixture of analytes to be determined and the analytical method used:

A: Extraction as received

B: Extraction after addition of phosphate buffer (pH 6–8)

C: Adjust pH 1–2 with sulphuric acid before extraction.

NOTE: Sample treatment, and extraction solvents are entered onto a worksheet.

A.2.2 PrecautionsMajor sources of error are contamination, in particular, phthalate esters from plastics. Teflon orglass containers should be used. All glassware should be solvent rinsed after use, then washed.After washing, contaminated glassware should be heated in a furnace to 400°C before re-use.

Pesticide grade solvents (or all-glass distilled solvent) are used to minimise interferences. Newbatches of solvents, reagents and reagent water must be checked for any interference problems.Reagent water is prepared using a Milli-Q system.

Solvents, such as dichloromethane can have toxic vapours whilst others such as ether, hexane andacetone are highly flammable. All extraction and concentration steps are to be conducted in a wellventilated fume cupboard.


Some reagents and pesticides may be carcinogenic. All samples and reagents should be treated aspotentially toxic. Neutralise and clean up any spills immediately. In addition to protective clothingand safety glasses, wear gloves when handling any suspected toxic materials.

A.2.3 ReagentsUnless otherwise specified, all reagents are of AR grade or equivalent. The calibration standards,check, spike and surrogate samples are prepared from certified reference materials/solutions.

A.2.3.1 ChemicalsSodium sulphate-granular (10–60 mesh) anhydrous: heat at 450°C for not less than four hours inshallow silica dishes before use.

Cotton wool (hexane washed): soak the cotton wool in a Schott bottle for 10 min with hexane,decant and rinse three times with hexane, then air dry.

NOTE: after washing, do not touch the cotton wool with hands or gloves, only use tweezers.

A.2.3.2 Reagent SolutionsPhosphate buffer: prepared from 1M HCl (75 mL) and 1M dipotassium hydrogen phosphate(125 mL), made up to 2,000 mL.

A.2.3.3 Standard SolutionsPrepare stock solutions from certified reference materials at a concentration of 200 µg/mL. Acetoneis generally used as the dilutant and the resultant solutions are stored in amber bottles with Teflonlined caps.

Prepare spike mixes and surrogates from the stock solutions and make up to volume with methanolor acetone as appropriate. Store all solutions in amber bottles with Teflon lined caps at or below4°C; in a flame-proof refrigerator.

A.2.4. ApparatusTable A.2.2List of preparative apparatus used

Apparatus Specifications

Separating Funnel Glass, fitted with a Teflon stopcock and a glass stopper. Capacity appropriate tothe sample volume to be extracted.

Filter Funnel Glass, 75–100 mm diameter with a hexane-washed cotton wool plug.

Graduated Tube 10 mL, glass with a B14 socket

Kuderna-Danish Flask 500 mL, with a B14 lower joint and a B19 top joint and fitted with hooks toattach (by springs) to the graduated tube.

3 Ball Snyder Column: Approx. 20 cm with B19 upper and lower joints.

Water Bath Six hole, with temperature control to ± 5°C.


A.2.5 Quality control protocol at sample extractionEach batch of samples includes: reagent blanks (10% of samples, minimum one) recovery blanksand recovery samples (5–10% of samples, minimum one). The samples for recovery data are spikedwith a mixture of certified standards appropriate to the samples being analysed. Similar mixturesof certified standards (10% of samples) are analysed with each batch of samples to cover the rangeand concentration levels expected in the samples tested.

Prior to sample extraction, the Section Head must record on liquid-liquid extraction worksheets theappropriate information. Ensure that all of the above quality control requirements are fulfilled.Enter onto the worksheet the spiking mixes and surrogates to be used with the volume andconcentration levels. Determine any special requirements of the client (analytes, detection levels)and/or analytical method, and record these details on the worksheet.

A.2.6 ProcedureThe following procedure is routinely followed for water quality samples. Any difference in thesample treatment or extraction solvent must be recorded on a worksheet.

1. Add a 500 mL aliquot of the water sample to the separating funnel. Prepare a batch as perworksheet list (to include reagent blanks, duplicates, replicates, recovery blanks and recoverysamples).

2. Depending on the analytical method, check the pH of the sample, adjust the pH or addphosphate buffer (25 mL) as indicated in Table A.2.1.

3. Add surrogate standards, if indicated on the worksheet, to all of the samples and spikingmixes.

4. Add dichloromethane (30 mL) to the separating funnel seal, then shake for two minutes withfrequent venting at the stopcock.

CAUTION: Vent immediately after the initial shake to prevent build-up of pressure.

5. Allow the two layers to separate and collect the organic layer in an Erlenmeyer flask. If anemulsion forms, and cannot be separated, transfer the organic phase plus the emulsion to asmaller separating funnel.

6. Repeat steps 4 and 5 twice and combine the three extracts. If an emulsion has formed, combineall extracts into a small separating funnel, remove the organic phase then add saturatedaqueous sodium chloride to break up the remaining emulsion. Add additionaldichloromethane to ensure quantitative extraction of the organic compounds.

7. Consult the analytical method to determine the extraction conditions for the relevant group ofanalytes. If a second extraction is required at a different pH, adjust the pH level and repeatsteps 4 to 6. Combine the three extracts.

8. Depending on the analytical method, extracts obtained at different pH values are combined ortreated separately. These modifications are noted in the procedures for individual methods.

9. Attach by springs (or clips), a 10 mL graduated tube to a 500 mL Kuderna-Danish (K-D) flask.Filter the organic extract collected in the Erlenmeyer flask (step 5) into the K-D flask through afilter funnel containing anhydrous sodium sulphate. Rinse the Erlenmeyer flask three timeswith 10 mL of dichloromethane into a filter funnel.

10. Add carborundum boiling chips, attach a Snyder Column which was prewet withdichloromethane (about 1 mL). Place on the water bath at 80–90°C and concentrate to about 5 mL.


11. For solvent exchange, leave the apparatus on the water bath, add 3 x 5 mL portions of theexchange solvent and concentrate to about 5 mL before each addition.

12. Remove from the water bath and allow to cool. Rinse the walls of the K-D with a minimalvolume of solvent.

13. If no further treatment is required before the analysis, adjust the volume in the graduated tubeto the required volume (usually 5 mL). This can be accomplished by the addition of solvent orremoval of the excess of solvent (evaporation with a stream of nitrogen over the surface). Thenormal procedure is to place a 1 mL aliquot into a vial for analysis and the remainder into asuitable amber vial with a Teflon lined cap. Store this reserve sample below 4°C.

14. If further clean-up or derivatisation is required, proceed from step 11 as required for theanalytes or class of analytes to be determined.


APPENDIX 3

Extraction of soil and foliage

A.3.1 Extraction procedure for endosulfan residues in soilA.3.1.1 NSW Agriculture, BCRIThe method employed for extraction of soil by NSW Agriculture Biological and Chemical ResearchInstitute (BCRI-120) is a modification of that of Sission et al. (1968) and de Faubert Maunder (1964).Organics are extracted by shaking with a solvent mixture of Nanograde dichloromethane andacetone (80:20, v/v). The coextractives are removed from the concentrated extract on an aluminacolumn and the eluate is chromatographed on a 30 m capillary column and quantitated by electroncapture detection.

A.3.1.2 University of SydneyThe method developed at the University of Sydney (CRDC project, US2C) involves the followingprocedure given in detail.

1. Thoroughly mix soil samples with a clean spatula before sub-sampling. Sieving of moistcracking soils to remove debris is impossible and drying to facilitate this would result in lossesof volatile chemicals such as endosulfan (Spencer and Cliath, 1975).

2. Weigh a 25 g sample accurately into a clean 250 mL conical flask sealed with a ground-glassstopper. Weigh accurately a further 10 g sample onto a watch-glass for moisture determinationby drying for two days at 110°C. Re-weigh and determine moisture content by difference.

3. Extract the soil in the conical flask with 150 mL of 25% acetone in dichloromethane byshaking on an orbital shaker at 165 rpm overnight (Braun Certomat provides reliability).

4. Decant the solvent carefully through a fluted filter paper containing 1–2 g of anhydrous Na2SO4

and collect 75.0 mL for analysis.

5. Transfer 75.0 mL to a Kuderna-Danish flask and evaporate to 5 mL.

6. Prepare a clean-up glass chromatography column using a cotton wool plug to retain 7 g ofalumina (7% H2O, w/w), overlaid with 1 g of anhydrous Na2SO4.

7. Apply the concentrated sample to the clean-up column and elute with 60 mL hexane followedby 60 mL 25% acetone in hexane. Collect and discard the first 10 mL of eluate (this fractioncontains principally DDE) and collect the remaining eluate in a Kuderna-Danish apparatus,ready for concentration.

8. Concentrate the eluate to less than 10 mL and rinse with hexane. (It is essential to ensure thatall dichloromethane is removed because it has a high electron capturing ability.)

9. Reduce the sample to less than 5 mL under a flow of nitrogen gas, transfer quantitatively to a10 mL volumetric flask and make up to volume with Nanograde hexane.

Alternatively, the evaporation steps using the Kuderna Danish apparatus may be replaced withTURBO-VAP equipment using automatic evaporation with heat and nitrogen gas to a specificvolume. Such TURBO-VAP equipment is now available including collection of evaporated solvent,avoiding the use of a fume cupboard with environmental dispersion into the atmosphere.


It should be noted that endosulfan lactone must be analysed independently and that endosulfandiol, although recovered well, is analysed with consistently low sensitivity, apparently as a resultof degradation on-column, even with capillary columns using low temperature injection withtemperature ramping. Fortunately, diol peak heights are constant with a given set of analyticalconditions, but a method involving derivatisation could be necessary to obtain comparablesensitivity.

A.3.1.2 Queensland Department of Primary IndustriesThis method describes the procedure for the extraction, identification, quantisation andconfirmation of the endosulfan and some related pesticides listed in Table A.3.1 in soil.

Table A.3.1Chlorinated Pesticides

p,p¢-DDEC14H9Cl4

Mol. Wt. 3191,1-dichloro-2,2-bis(4-chlorophenyl) ethylene

p,p¢-DDTC14H9Cl5

Mol. Wt. 354.5l,l,l-trichloro-2,2-bis (4-chlorophenyl)ethane

o,p¢-DDTC14H9Cl5

Mol. Wt. 354.5l,l,l-trichloro-2(2-chlorophenyl)-2(4-chlorophenyl) ethane

α-EndosulfanC9H6Cl6O3SMol. Wt. 406.96,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro-6,9-methano-2,4,3-benzo[e]-dioxathiepin-3-oxide

β-EndosulfanC9H6Cl6O3SMol. Wt. 406.96,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro-6,9-methano-2,4,3-benzo[e]-dioxathiepin-3-oxide

CI CI

CCI2

C

CCI3

CH CICI

CI

CI

CCI3

CH

CI

CI

CI

CI

CICI

O

O

SO

CI

CI

CI

CI

CICI

O

O

SO


Endosulfan sulphateC9H6Cl6O4SMol. Wt. 422.96,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro-6,9-methano-2,4,3-benzodioxathiepin 3,3-dioxide

A.3.1.3.1 Referenced documentsThe following document is referred to in this method: Pesticide Chemistry Quality Manual

A.3.1.3.2 DefinitionsThe ‘Detection Limit’ is the lowest level of an individual compound. To quantify that the methodcan confidently confirm at least the presence of an individual compound.

Detection Limit for Method: 0.002 mg/kg (generally).

This figure is a guide only and varies from one pesticide to another as well as from sample tosample and depending on the manner of reporting.

A.3.1.3.3 PrincipleThe sample is extracted with methanol/water. The pesticides are extracted into n-hexane fromaqueous methanol. The hexane phase is dried over sodium sulphate and cleaned up on a Florisilcolumn. The chlorinated pesticides present in the resultant concentrate are determined by GLC andidentity confirmed by GC-MS.

A.3.1.3.4 ReagentsSolvents: Methanol, n-hexane, diethyl ether, iso-octane

Chemicals: Deionised water, Sodium chloride (LR grade) in deionised water (saturated solution).Whatman glass microfibre filters (GF/A) (9.0cm). Anhydrous granular sodium sulphate (AR Grade),Florisil. (60–100 mesh P.R.)

Standard solutions: The following standard solutions are prepared and stored in accordance withprocedures described in the Pesticide Chemistry Quality Manual.

1. Bulk Solutions: Prepare individual stock solutions in iso-octane of each of the followingpesticides at a concentration of 200 mg/L in 50 mL volumetric flasks.

a-endosulfan p,p´-DDE trifluralin.

b-endosulfan o,p´-DDT

endosulfan sulphate p,p´-DDT

2. Intermediate solution: Prepare the following composite solutions in iso-octane at aconcentration of 2 mg/L of each standard. All intermediate solutions are stored in a freezer anddiscarded every six months.

OC-1: a-endosulfan p,p´-DDE trifluralin.

b-endosulfan o,p´-DDT

endosulfan sulphate p,p´-DDT

CICI

CI

CICI

O

S

O

OO

CI


3. Working solutions: Prepare working solutions from the intermediate standards at aconcentration of 0.02, 0.10 and 0.2 mg/L of each standard material in hexane. Workingsolutions are stored in a refrigerator. Fresh standards are prepared and compared to the oldstandards before the old standards are discarded.

A.3.1.3.5 Apparatus500ml Quick-fit Erlenmeyer flasks, Vacuum filtration apparatus, Rotary evaporator, Water bath, Gaschromatograph/EC detector, Glass column (OV-1), 500ml round bottom flasks, 500ml separatingfunnels, 100ml stoppered measuring cylinders.

A.3.1.3.6. Sampling and sample preparationRefer to the Pesticide Chemistry Quality Manual.

A.3.1.3.7. Procedure

Preliminary screenBecause of the possibility that some soil samples could contain very high levels of pesticideresidues, a preliminary screening is advisable for samples with no background information tominimise the possibility of contamination of the laboratory and/or of the analytical apparatus.Before commencing the analysis proper, weigh soil (5 g) into a stoppered test tube. Add hexane(10mL). Insert a soni-probe for three minutes and then filter a small quantity through glass fibrepaper. Assess the likely level in the soil by GLC with a Hall detector.

Moisture contentIn general, most soil samples submitted for analysis are analysed and reported on a ‘dry weight’basis. In such cases, weigh sample (20 g) into a tared moisture dish. Dry overnight at 105°C. Allowto cool in a desiccator and reweigh. Calculate the % moisture.

Residue analysisAll soil samples are air dried at 400C for 48 hours and ground to a particle size, no greater than,2mm.

Weigh sample (25 g) into a Quickfit conical flask (500 mL). Add methanol (100 mL) and water (25mL). Record the volume of water added. Stopper and shake for one hour on a wrist-action shaker.Filter under vacuum through glass fibre filter paper and record the volume. Transfer the extract to aseparating funnel (500 mL) and add hexane (50 mL).

Shake vigorously for two minutes and add sodium chloride solution (10 mL) and water (300 mL).Shake the mixture gently, but thoroughly, for two minutes and allow to settle.

Discard the aqueous layer after the layers separate. Transfer the hexane solution to a stopperedmeasuring cylinder (100 mL) and record the volume. Add sodium sulphate (» 5 g) and shake. Allowto stand for at least one hour to dry the hexane.

Prepare a Florisil column by packing a 2.0 cm i.d. column (fitted with a fritted glass disc and tap) toa height of 9 cm with Florisil (» 25 g) which has been dried at 105°C. Introduce a layer of sodiumsulphate (» 1 cm) to cover the Florisil. Wash the column with hexane (50 mL) before use.


Transfer the sample solution to the column and allow to run at two drops/sec into a round-bottomed flask (500 mL). Elute (See note) the column with:

(a) ether/hexane (20+80 v/v) (100 mL) and

(b) ether/hexane (60+40v/v)(150mL) Collect the eluate separately and evaporate to »5mL on arotary evaporator with water bath at 40°C. Transfer the concentrate to separate volumetricflasks (10ml) for determination.

NOTE

The second eluate [(60+40v/v) ether/hexane], contains any b-endosulfan and endosulfan sulphate.

A.3.2 Extraction of foliageThe extraction of foliage can be made by blending leaves with a mixture of dichloromethane andacetone (80:20) in an omnimixer. Extraction efficiencies should be established by spiking foliagesamples at four different levels, one very close but slightly above the detention limit of the methodwith replication to establish the variation of the method. This method is currently in use at theBCRI.

Another method is extensively adapted from one developed by Chopra and Mahfong (1977) forextraction of endosulfan and metabolites from tobacco leaves. It has been employed for study of therate of dissipation of endosulfan from cotton leaves (Coleman and Kennedy, 1993).

The steps of the procedure are:

1. Chopped leaves (ca. 15 g fresh weight) with 20 g Na2SO4 plus 5 g NaCl are homogenised in50 mL of 25% (v/v) acetone in hexane (Nanograde) in a Sorvall Omnimixer equipped with a250 mL cup for one minute (‘spikes’ can be added at this stage). The solvent is filtered through125 mm filter paper containing 6 g anhydrous Na2SO4 into a 100 mL round-bottom flask. Thehomogeniser cup is rinsed with about 20 mL of 25% (v/v) acetone-hexane, with the plantmaterial in the filter.

2. After evaporation to 5 mL, the sample is transferred quantitatively to a 10 mL volumetric flaskand made up to volume with hexane.

4. Clean-up is achieved on a glass column (7 mm ID) containing 1.90 g Florisil overlaid with2–3 cm anhydrous Na

2SO

4. The packed column is rinsed with 10 mL hexane and 5 mL of fresh

hexane placed directly in a 50 mL pear-shaped flask placed under the column. 1 mL of the leafextract is loaded onto the column and allowed to settle into the Na2SO4. 30 mL of diethyl etheris flushed through the column.

5. The material in the flask is evaporated to 2–3 mL (rotary vacuum evaporator, eg. Buchi orunder a flow of N2 ) and transferred quantitatively to a 5 mL volumetric flask and made up tovolume with hexane. The solution may be diluted with hexane if necessary for gaschromatographic injection (see below).

This treatment results in removal of most of the plant pigments, but some yellow colour mayremain, depending on the carotenoid content of leaves. For samples that will be undiluted beforegas chromatographic analysis, additional clean-up from pigments may be obtained by use of a12 cm column containing 2.50 g Florisil, with about 5–10% less recovery of endosulfan andmetabolic products. The procedure provides better than 85% recovery for endosulfan isomers andall breakdown products.


APPENDIX 4

Gas chromatographic analysis

A.4.1. Determination of pesticides using gaschromatography—mass spectrometryA.4.1.1 PrefaceThis method (WELOR301) was provided by NSWDLWC and is based on USEPA Method 625,‘Bases/neutrals and acids’. The procedure can be used for the determination of a range of organicclasses, including pesticides in complex matrices such as soils, sludges and waste water. It can alsobe used to complement or to confirm results using a more selective detector. The range of chemicalsinclude the organochlorines and other pesticides which are of concern to the Department. Thesepesticides are generally used in the agricultural areas of NSW where cotton and rice are harvested.

Modifications to this method are:

a) The primary GC column is a 30m x 0.25mm x 0.25µm DB5 column and a mass selective (MS)detector.

b) Range of analytes.

A.4.1.2 ScopeThis procedure has been validated for the pesticides listed in Table A.4.1. The laboratory maintainscertified reference materials/standard solutions for the analytes reported. The method is applicableto the determination of pesticides in soils, sediments, sludges and waste water and iscomplementary to the conventional methods for the determination of pesticides in groundwater,surface water, potable water, and marine water. Other classes of organic compounds for which thismethod is also applicable are the phenols (WELOR302), polycyclic aromatic hydrocarbons(WELOR303), petroleum hydrocarbons (WELOR304) and organic acid herbicides (WELOR305).

A.4.1.3 PrinciplesSamples are extracted using an appropriate method (WELOR101,102). The dried extracts areconcentrated using a Kuderna-Danish flask and the solvent is exchanged to hexane (or a suitablesolvent). The final volume is initially equivalent to a 1:100 fold concentration for water samplesand a 1:20 fold concentration for soil samples. Further concentration or dilution is dependent onthe level of detection required.


A.4.2 Determination of chlorinated pesticidesChlorinated pesticides are determined by Electron Capture Detector using an OV-1 GLC column(injection volume 1–5mL).

Instrument Parameters (Varian 3600)

Column (1): 3% OV-1 on Chromosorb WHP 100/120 mesh (7' 5" x 2mm i.d.).

GLC Conditions Column (1)

(Varian 3600) (3% OV-1)

Column Temp. 195°CInjector Temp. 230°CCarrier Gas Nitrogen

Carrier Gas(flow rate) 25 mL/min

Detector ECD

Column Pressure 7.2 kPa

Detector Oven Temp 350°C

A.4.2.1 Calculations:

(a) The concentration of the pesticide in the sample (on a dry weight basis) is—

level(mg/kg) = C x A1 x (W1 + 125) x 50

W x A2 x V1 x V2

where C = concentration of the standard (mg/mL)

A1 = peak area of the sample

A2 = peak area of the standard

W1 = weight (in g) of water contained by the test portion of soil

V1 = recorded volume of the methanol/water filtrate

V2 = recorded volume of n-hexane recovered

W = weight (in g) of the test portion of air dried soil

assuming 100 mL of methanol and 25 mL of water are used to extract the soil sample, and 50 mL ofn-hexane is used in the partitioning step.

A.4.2.2 Reporting of results:Examples of the various formats of test reports used are given in the Pesticide Chemistry QualityManual.


Table A.4.1List of pesticides for which WELOR301 is applicable.

Compound Retention time Primary Secondary Group Detection Limit

(min) ion ion *(ng/L)

Dichlorovos 9.73 109 TI 1 100

Molinate 13.86 126 187 1 50

Demeton-S-methyl 14.91 88 60 2 100

Trifluralin 15.62 306 264 3 50

Phorate 15.83 75 121 3 100

a-BHC 15.95 219 181 3 & 4 100

Thiometon 16.00 88 125 4 100

Dimethoate 16.22 87 93 4 100

Atrazine 16.45 200 215 4 50

Hexachorobenzene 16.55 284 286 4 50

b-BHC 16.57 219 181 4 100

d-BHC 16.73 219 181 4 100

Diazinon 17.12 179 304 4 100

Disulfoton 17.25 88 90 4 100

g-BHYC 17.30 219 181 4 100

Parathion methyl 18.39 263 109 5 100

Heptachlor 18.67 272 274 6 200

Prometryn 18.72 241 226 6 100

Bromacil 19.25 205 207 6 100

Malathion 19.54 125 173 7 200

Aldrin 19.73 66 263 7 100

Chlorpyrifos ethyl 19.90 314 316 7 100

Metolachlor 19.90 162 238 7 100

Pendimethalin 21.00 252 281 8 100

a-Chlordane 21.90 373 375 9 50

a-Endosulfan 22.38 237 265 9 200

g-Chlordane 22.50 373 375 9 50

p,p-DDE 23.33 246 318 10 100

Dieldrin 23.40 79 263 10 200

Oxyfluorofen 23.75 300 252 10 100

Fluazifop butyl 24.43 282 383 11 200

Endrin 24.53 67 345 11 400

b-Endosulfan 24.66 237 265 12 200

p,p-DDD 25.09 235 165 13 200

Endrin Aldehyde 25.48 345 67 13 400

Heptachlorepoxide 20.90 353 357 7 100

Parathion ethyl 21.54 109 291 8 100

Sulprofos 25.84 156 322 13 100

Endosulfan Sulphate 26.55 235 272 14 200

p,p-DDT 26.78 235 237 14 100

Haloxyfop 28.24 302 433 15 100

* The detection limits were determined from reference standards and not recovery samples.


A.4.2.3 Confirmation of reportable residues:GC-MS Confirmation: Results determined by GLC with EC detection may be confirmed by GC-MS.

Use splitless 1 or 2mL injections. For the MS method, use the SIM mode for48 ions in three time groups, representing 21 compounds (3 ions percompound).

GLC column J & W Scientific DB-1 capillary, 30 m x 0.25mm i.d. x 0.25 mm.

Temperatures:

Injector 250°CColumn 60°C for 2 min, 60°C to 240°C @ 30°C/min

240°C for 10 min

Interface 280°C

Precision: Information on detector linearity is to be included.

Method Recovery: Information on the recovery of organochlorine residues from soil is to beincluded.

A.4.3 Determination of nitrogen and phosphorus containingchemicals using gas chromatographyA.4.3.1 PrefaceThis method (NSWDLWC WELOR204) is based on USEPA Method 507 for the determination ofnitrogen and phosphorus containing pesticides. A gas chromatograph fitted with a nitrogen-phosphorus detector is used.

Modifications from this method are:

a) a 500 mL aliquot volume of sample is extracted

b) methyl tert-butyl ether is replaced by hexane as the exchange solvent

c) primary GC column is 30m x 0.25mm x 0.25µm DB1 column. Confirmation is by use of a 30mx 0.25mm x 0.25µm DB5 column and a mass selective (MS) detector, or a 30m x 0.25µm DB1301 column and a flame photometric (FP) detector for phosphorus/sulphur containingchemicals.

A.4.3.2 ScopeThis procedure has been validated for the analytes listed in Table A.4.2. The laboratory maintainscertified reference materials/standard solutions for all analytes reported. The method is applicableto the determination of nitrogen and phosphorus containing chemicals in groundwater, surface,potable, marine and waste waters. Soil and sediment samples can be analysed by this method, or bythe use of selective detection (WELOR301).

A.4.3.3 Principles

The aqueous sample is extracted with dichloromethane in a separating funnel. The dried extract isconcentrated using a Kuderna-Danish flask and solvent exchanged to hexane (or suitable solvent) toa final volume equivalent to a 1:100 fold concentration.


The final extract is analysed by gas chromatography using a nitrogen phosphorus detector.Confirmation is carried out by GC with MS detection, or by the use of a GCV column of differentpolarity with an FP detector.

Table A.4.2Nitrogen and phosphorus containing pesticides forwhich WELOR204 is applicable.

Compounda CAS Number Method Detection Acceptable Recovery

Limit (µg/L) Range (%)

Acephate 30560–19–1 0.1 70–130d

Atrazine 1912–24–9 0.1 62–122b

Bromacil 314–40–9 0.1 61–121b

Diazinon 333–41–5 0.1 85–145b

Dichlorvos 62–73–7 0.1 67–127b

Disulfoton 298–04–4 0.1 59–119b

Demeton-S-methyl 919–86–8 0.1 75–135b

Chlorpyrifos-ethyl 2921–88–2 0.1 62–122c

Metolachlor 51218–45–2 0.1 63–123b

Malathion 121–75–5 0.5 40–115c

Parathion-methyl 298–00–0 0.1 70–130d

Parathion-ethyl 56–38–2 0.1 70–130d

Phorate 298–02–2 0.1 70–130d

Prometryn 41198–08–7 0.1 63–123b

Molinate 7287–19–6 0.1 68–128b

Trifluralin 2212–67–1 0.1 51–140c

Pendimethalin 1582–09–8 0.1 59–119c

Sulprofos 40487–08–7 0.1 68–128c

Profenofos 41198–42–1 0.1 57–117c

Dimethoate 35400–43–2 0.1 70–130d

Thiometon 60–51–5 0.1 70–130d

a Certified reference standards are maintained for all analytes.b Range obtained using procedure described in USEPA Method 507, (Section 10.3.2. plus results from Table A.4.2). Range quoted

is Mean ± 30% or Mean ± 3 x S.D. (whichever is the larger).c Range obtained from results in this laboratory.d When no recovery data are available, the recovery range used is 70–130.


A.4.4 Analysis of endosulfan using gas chromatographyA.4.4.1 PrefaceThe following protocol was submitted by the Gatton College Analytical Laboratory, University ofQueensland, where studies on pesticide drift as part of the overall program are being conducted(contact, Nicholas Wood). This is modified from the USEPA Method 608, to include use of moremodern capillary columns. It is reproduced here in full, although it is expected that individuallaboratories will use procedures differing slightly in details such as choice of capillary column. It isanticipated that all laboratories concerned in the program will employ capillary columns forroutine analyses, however, in case a packed column method is required, an effective columnpacking previously used at the University of Sydney is SE-30 (5%):DC-200 (5%) on Gaschrom Q,80–100 mesh, a liquid phase allowing separation of endosulfan isomers, endosulfan sulphate andthe full range of degradation products (Guerin et al., 1993). The identity of all peaks will be verifiedby inclusion of some analyses on columns with a different separation pattern and using internal orexternal standards for all compounds separated. Alternatively, as indicated here, massspectrometric verification of peaks can be performed.

USEPA Method 608 prescribes packed columns. The following procedure recommended byHewlett-Packard (Klee, 1989) substitutes a high resolution capillary column without compromisingthe specified criteria for EPA-608.

A.4.4.2 Equipment and conditionsInstrument: Hewlett-Packard 5890A gas chromatograph with HP7673A

automatic injector and an HP 3396A integrator.

Detector: Electron capture detector with N2 make-up gas.

Detector temperature: 325°C

Carrier gas: CIG Ultra-High Purity Helium, with indicating moisture andoxygen traps.

Injector port temperature: 190°C

Injection mode: Splitless

HP 5890A split-vent flow rate: 50 mL/min

HP 5890A split-vent On time: 1 minute

Column: 25 m x 0.2 mm, 0.11 um Ultra-2 capillary column (5% phenylmethyl silicone, part no.19091B-002) at a column head pressureof 30 psi (to maximise effective resolution.

Any other column giving good resolution (see section 4.1–4.2).

Injection volume: 0.5 mL

Standards: Supelco (Bellefonte) or other suppliers and diluted in iso-octaneto concentrations of 0.5–3.0 mg/mL.

Analysis time: For EPA-608 pesticides, 18 min.


A.4.4.3 Optimised GC analysis conditionsMultiple temperature ramp: 80–175°C @ 30°C/min

175–225°C @ 2.5°C/min

225–275°C @ 10°C/min

This gives optimum separation of endosulfan isomers from other EPA-608 pesticides. The startingtemperature of 80°C is a compromise between stationary-phase focussing and a short analysis time.In this protocol, there should be no ‘splitless effect’ and sample volume, solvent type and startingoven temperature are not critical.

A.4.4.4 Confirmatory analysisA separate Hewlett Packard bench-top GC/MS (5890 Series II coupled to a 5790 MSD) will be usedas a confirmatory tool.

A.4.4.5 Extraction solventSolvents to be tested to extract metal and inert surfaces (targets) include isopropanol and iso-octane(hexane should be suitable for endosulfan). Previously, ethylene digol has been used to removespray drift dyes from leaf surfaces. A solvent of the type discussed in Section 2.3 should provesuperior for this purpose with endosulfan isomers, which can be expected to be absorbed into theleaf tissue.


APPENDIX 5

Immunoassay of endosulfan and diuron

A.5.1 Endosulfan immunoassayAntibodies were raised as described (Lee et al., 1994).

A.5.1.1 Preparation of endosulfan standardA stock solution of 100 ppm endosulfan (isomer mixture: 70% alpha and 30% beta) is prepared inmethanol. From this stock solution, 100 ppb is prepared by 1/1,000 dilution in purified water andthen serially diluted to obtain 100 ppb, 30 ppb, 10 ppb, 3 ppb, 1 ppb, 0.3 ppb and 0.1 ppb inborosilicate glass tubes for the standard curve.

A.5.1.2 Laboratory analysis ELISATo each antibody-coated plate, 100 µL of endosulfan standard or sample followed by 100 µLHRP-conjugate diluted in PBS or PBS with 0.5% fish gelatin (w/v), are added and incubated for onehour at 20°C. After this incubation, plate contents are removed and the plate is then washed indistilled water. Hydrogen peroxide substrate/chromogen (3,3’,5,5”-tetramethylbenzidine-hydrogenperoxide in acetate buffer, pH 5.5, 150 µL) is added and incubated 30 minutes at 20°C. The colourdevelopment is stopped by adding 50 µL 1.25M sulfuric acid, and the plate is read at 450 nm.

A.5.1.3 Field analysis ELISAFour drops (160 µL) of sample and four drops (160 µL) of HRP-conjugate are added to a tube andincubated for 10 minutes at ambient temperature. The tube is washed with tap water four timesand shaken vigorously. Then four drops of substrate (300 mL) followed by four drops of chromogen(150 µL) are added to a tube for colour development. After five minutes, four drops of stoppingsolution (0.625M sulfuric acid) are added, and the absorbance is read with a portable photometer.

A.5.2 Diuron immunoassayA.5.2.1 Preparation of Diuron StandardA stock solution of 100 ppm diuron is prepared in methanol. From this stock solution, 100 ppb isprepared by 1/1,000 dilution in purified water and then serially diluted to obtain 5 ppb, 1 ppb,0.5 ppb, 0.1 ppb, and 0.05 ppb in borosilicate glass tubes for the standard curve.

A.5.2.2 Laboratory and field analysis ELISABoth laboratory and field analysis for diuron are exactly the same as for endosulfan (see above),using diuron antibody-coated plates and diuron-HRP conjugate.


References

Australian Standard 2031.1—1986. Selection of containers and preservation of water samples for chemical andmicrobiological analysis. Part 1—Chemical. Standards Association of Australia, Standards House, 80 ArthurSt., North Sydney, N.S.W.

Changes in Official Methods of Analysis, (1968), J.A.O.A.C. 51, 472–481, paragraphs 24.222–24.227.

Chopra, N.M. and Mahfouz, A.M. (1977) Metabolism of endosulfan I, endosulfan II and endosulfan sulfate intobacco leaf. J. Agric. Food Chem. 25, 32–36.

Coleman, S. (1993) The fate of endosulfan applied to hydroponically-grown cotton plants. Honours thesis,Department of Agricultural Chemistry and Soil Science, The University of Sydney.

de Faubert Maunder, M.J., Egan, H., Godly, E.W., Hammond, E.W. and Roburn, J. (1964) Clean-up of animal fats anddairy products for the analysis of chlorinated pesticide residues. The Analyst 89, 168–174.

Guerin, T.F. and Kennedy, I.R. (1992) Distribution and dissipation of endosulfan and related cyclodienes in sterileaqueous systems: Implications for studies on biodegradation. J. Agric. Food Chem. 40, 2315–2323.

Guerin, T.F., Kimber, S.W.L. and Kennedy, I.R. (1992) Efficient one-step method for the extraction of cyclodienepesticides from aqueous media and the analysis of their metabolites. J. Agric. Food Chem. 40, 2309–2314.

Kennedy IR, Harris CR, Sánchez-Bayo F, Kimber SWL, Southan S, Hugo L, Caldwell RA, Shivaramaiah H, Feng L,van Zwieten L, Lee NJ, Beasley H, Wang S & Skerritt JH (1997) Cotton pesticides in perspective. ISBN 1-86451-306-3. Discussion Paper, Vol 1, 73 pp. CRC for Sustainable Cotton Production, Australian Cotton ResearchInstitute, Narrabri.

Kimber, S.W.L., Ahmad, N. and Kennedy, I.R. (1994) The fate of endosulfan sprayed on cotton. Proc. World CottonConference. Brisbane, February 14–17/1994.

Klee, M.S. (1989) Optimised EPA-608 analysis using an ultra-2 high resolution capillary column, Hewlett PackardApplication Note No. 228–95.

Kovacs, M.F. (1963) Thin-layer chromatography for chlorinated pesticide residue analysis. J.A.O.A.C., 46, 884–893.

Kovacs, M.F. (1965) Thin-layer chromatography for pesticide residue analysis. J.A.O.A.C., 48, 1018–1022.

Lee N, Beasley HL, Kimber SWL, Silburn M, Woods N, Skerritt JH & Kennedy IR (1997) Application ofimmunoassays to studies of the environmental fate of endosulfan. J. Agric. Food Chem., 45: 4147–4155

Lee, N., Skerritt, J.H., Kennedy, I.R., McAdam, D.P., Beasley, H.L. and Hill, A.S. (1994) Development of laboratoryand field ELISA tests for detection of endosulfan in water and soil. Proc. World Cotton Conference. Brisbane,February 14–17/1994.

Mills, P.A. (1959) Detection of semiquantitative estimation of chlorinated organic pesticide residues in food bypaper chromatography. J.A.O.A.C., 42, 734–740.

Pesticides Analytical Manual, (1972), U.S. Dept. of Health, Education and Welfare, 1, Sections 211.14 d, 202, 212,410 and 411.

Peterson, S.M. and Batley, G.E. (1993) The fate of endosulfan in aquatic ecosystems. Environmental Pollution 82,143–152.

Sission, D.J. and Telling, G.M. (1968) J. Chromatog. 33, 435–449

Spencer, W.F. and Cliath, M.M. (1975) Vaporisation of chemicals. In: Environmental Dynamics of Pesticides, Haque,R. and Freed, V.H., eds. Plenum Press.

USEPA Method 608—Organochlorine pesticides and PCB’s, Federal Register, Vol. 49, No. 209, October 1984.

quality assurance in pesticide sampling and analysis · made in sampling and analytical procedures...

Documents